We recently had problems with vsphere-csi-driver pods going in CrashloopBackoff state because the default probe-timeout is too short. We created a PR: https://github.com/kubernetes-sigs/vsphere-csi-driver/pull/2997. Taking a step back, we found that this timeout is increased to 3s or more in a lot of places: See [here](https://github.com/search?q=org%3Akubernetes-sigs+probe-timeout%3D&type=code) and [here](https://github.com/search?q=org%3Akubernetes-csi+probe-timeout%3D&type=code). Maybe the default should be increased to 3 or 4s?